Analyst, Janztayy, 1979, Vol. 104, pp. 35-40 35 Determination of Germanium, Arsenic, Selenium, Tin and Antimony in Complex Samples by Hydride Generation - Microwave-induced Plasma Atomic-emission Spectrometry Wayne B. Robbins and Joseph A. Caruso and Fred L. Fricke Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 4522 1, USA Cincinnati District Food and Drug Administration, 1141 Central Parkway, Cincinnati, OJzio 45202, USA This paper describes the utilisation of a semi-automatic hydride generation device coupled to a microwave-induced argon - helium plasma that is used with a 0.5-m monochromator for the analysis of germanium, arsenic, selenium, tin and antimony in several complex samples. Incorporated in the system are a condensation tube and a Chromosorb 102 column that are used to separate the analyte species from hydrogen evolved during the course of the generation reaction, and to separate the analytes from condensed contaminants that cause spectral background interferences. Results reported are standard recoveries and relative precision for ger- manium, arsenic, selenium, tin and antimony in whole blood and in enriched flour. Also reported are arsenic and antimony values for NBS orchard leaves.Results obtained on complex samples are compared with the precision obtained for aqueous standards. Keywords : Hydride generation ; microwave-induced filasnza ; atomic-absorption spectrometry ; germanium, arsenic, selenium, tin and antimony determination The utilisation of hydride generation in atomic-spectrometric analysis is continually increasing in p~pularity.l-~ This type of analysis allows the separation and pre-concentration of the analyte from sample matrices.Elimination of the sample matrix serves to circumvent the problem of chemical interferences in various emission sources and atom reservoirs that are a frequent complication in techniques utilising solution nebulisation. The resulting advantages of hydride generation include lower detection limits and extended linear ranges. Fricke et aL6 described the coupling of a semi-automatic hydride generator to a microwave-induced argon - helium plasma via a condensation tube immersed in liquid nitrogen and chromatographic separation of background interferences on a Chromosorb 102 column. Advantages over the conventional hydride generation - atomic-absorption systems included extension of linear ranges, elimination of the use of hollow-cathode lamps, and adaptability to multi-element analyses.Linearity of up to three orders of magnitude, and detection limits in the sub parts per billion (parts per lo9) range were reported for germanium, arsenic, selenium, tin and ant imony.6 As complex samples were not investigated in that study, it was believed that the applica- bility of the system could be established more clearly by investigation of several complex samples. This was the purpose of the work described in this paper. Experimental Apparatus A 2450-MHz microwave generator and Eversen &wave cavity were supplied by Opthos Instrument Co. A silica plasma containment tube (110 x 1.7 mm id.) was supplied by Amersil Inc.The monochromator was a Jarrell-Ash 0.5-m instrument with a 1180 lines mm-l grating blazed for 300 nm, with a reciprocal linear dispersion of 1.6 nm mm-l in the first order. A Jarrell-Ash MVAA 82500 atomic-absorption spectrometer was used in the emission mode with a 100-pm straight-edge fixed slit for the entrance slit and a 1P 28 photomultiplier36 Analyst, Vol. 104 tube mounted at a 150-pm exit slit as the detector. The recorder was a Hewlett Packard, Model 7107 B. The chromatograph was a length of Polypenco Nylaflow pressure tubing, 3ft x 4.7 mm i.d., packed with Chromosorb 102, 60-80 mesh. The condensation tube was a siliconised glass tube of 1.5 mm i.d. with a 40-cm length packed with 1.0-mm siliconised glass helices.The hydride generator was as described by Fiorino et al.2 A schematic diagram of the instrumental arrangement is shown in Fig. 1 and a discussion of the development of the system has been published elsewhere.6 ROBBINS et al.: DETERMINATION OF GERMANIUM, ARSENIC, I - . - , Microwave Plasma generator - containment Monochromator PMT 1 Argon column tube - CaC12 .2H20 Vent to d-. Three-way valve, V2 Condensation - 4- Three-way valve, V1 a Helium CaC12 Q I Hydride I generator Electronics EI- 1 Recorder I I ~ Fig. 1. Block diagram of the instrumental arrangement. Reagents Stock solutions of Se4+, Sn4+ and Sb3+ were prepared from Fisher atomic-absorption standards while Ge4+ and As3+ solutions were prepared from Ventron [and E.M. Laboratories standard stock solutions, respectively.A 4% m/V solution of sodium tetrahydroborate(II1) stabilised with 1% m/V of sodium hydroxide was used to generate the hydrides from solutions in hydrochloric acid. Hydrochloric acid. Reagent grade. Arg0.n. Commercial grade, 99.998%. Helium. Calcium chloride. dihydrate. Acid mixture. acids supplied by G. Frederick Smith Chemical Co., Columbus, Ohio. Commercial grade, 99.999%, passed through powdered calcium chloride. Fisher ACS-grade %mesh granular and Baker ACS-grade powdered Nitric - sulphuric - perchloric acids (4 + 4 + l), prepared from re-distilled Instrumental Conditions power and 0 W reflected power. respectively. The microwave-induced argon - helium plasma (MIP) was operated at 110 W forward Argon and helium flow-rates were 400 and 300 ml min-1,January, 1979 TIN AND ANTIMONY I N COMPLEX SAMPLES 37 The hydride generator was operated as described by Fiorino et aL2 This device serves to meter a precise amount of sodium tetrahydroborate( 111) solution into the acidic mixture of analyte through solenoid valves controlled by a timer.Solutions are transferred by means of carefully controlled gas pressure. The chromatographic column was maintained at room temperature (approximately 23 "C) . Deviations of &3 "C did not significantly affect the chromatographic characteristics of the system. Analytical wavelengths used during this work were germanium 303.9, arsenic 193.7, selenium 196.0, tin 317.5 and antimony 259.8 nm. Preparation of Samples Whole blood A 5.0-ml volume of expired whole blood obtained from the Paul I.Hoxworth Blood Bank, Cincinnati, Ohio, was mixed with 30 ml of the acid mixture and was heated until fumes of sulphur trioxide were evolved. The final acid concentration was adjusted to that deter- mined as described under Optimisation of Signal with hydrochloric acid. Enriched $our Flour (2.00 g) that had been dried in an oven overnight at 85 "C was mixed with 30 ml of the acid mixture and heated until fumes of sulphur trioxide were evolved. The final acid concentration was adjusted to that determined as described under Optimisation of Signal with hydrochloric acid. NBS orchard leaves 1571 Portions of approximately 1.0 and 1.5 g for arsenic and antimony, respectively, that had been dried as specified by NBS, were accurately weighed, mixed with 30ml of the acid mixture and heated until fumes of sulphur trioxide were evolved.The final acid concentra- tion was adjusted to that determined as described under Optimisation of Signal with hydro- chloric acid. Owing to possible positive or inhibitive effects in the quantitative generation of the hydrides because of the presence of various salts in the digestion matrix, a standard additions graph was made to determine the concentration in each instance, rather than a determination from a calibration graph. Although no attempt was made to determine the final oxidation state of the elements after the digestion procedure, the recoveries from the spiked samples indicate clearly that the oxidation states present are converted into the hydrides with no apparent difficulty.Wet as opposed to dry oxidation was chosen as it presented fewer problems when all five elements were taken into consideration. Determination of Germanium, Arsenic, Selenium, Tin and Antimony With reference to Fig. 1, the condensation tube was isolated from the reaction chamber by adjusting the three-way valve V1 and then isolated from the vent leading to the hood and the chromatographic column by adjusting the second three-way valve V2. The con- densation tube was cooled in liquid nitrogen for 2.5 min, then V1 and V2 were adjusted so that the condensation tube was open on one end to the reaction chamber and on the other end to the hood vent. A 20-ml aliquot of the solution containing the element of interest was placed in the reaction tube, the tube attached to the head of the hydride generator and the timer activated. The resulting reaction swept the hydrides through granular calcium chloride desiccant into the condensation tube where they were condensed, together with any volatile contaminants, while hydrogen was vented to the hood.Valve V1 was then adjusted to close the condensation tube to the reaction chamber and allow helium to pass through to the hood for 0.5 min; this operation removed any residual hydrogen not vented during the course of the reaction. Valve V2 was opened to the chromatographic column and the con- densation tube was immediately immersed in a water-bath at 80 "C. The helium flow then caused the vaporis2d mixture of hydrides and other condensed contaminants to pass through powdered calcium chloride desiccant into the chromatographic column, where the hydrides and contaminants were separated, and finally into the MIP where the elemental excitation took place.Areas under the peaks obtained with the strip-chart recorder were used to represent the signal, and were measured with a planimeter.38 ROBBINS et at?. : DETERMINATION OF GERMANIUM, ARSENIC, Analyst, T/d. 104 Optimisation of Signal Owing to the nature of the contaminants in the sodium tetrahydroborate(II1) and its varying quality, depending upon the manufacturer and the batch, it is necessary to optimise the reaction conditions needed to produce the elemental hydrides each time a different batch or a different manufacturer's material is used. To do this, the concentrations of the acid solutions from which the hydrides were generated were varied until a maximum signal was achieved.The hydrides of germanium, arsenic, tin and antimony are optimally generated from solutions of the same acid concentration (1.2-2.4 N), while the selenium hydride is optimally generated from more acidic solutions than the other four hydrides. The wavelength for each analytical line was optimised with the analyte signal. Standard Recoveries of the desired element to the sample and acid mixture prior to digestion. calculated by comparing peak areas with those obtained from spiked samples. Recoveries were measured by adding several Inicrolitres of a concentrated aqueous solution Recoveries were Spiked Samples Owing to possible positive or inhibitive effects in the generation of the hydrides, because of the presence of various salts in the digestion matrix, standard solutions were added to a sample after the digestion process was complete, in order to produce the spiked sample.Results and Discussion Standard recoveries for germanium, arsenic, selenium, tin and antimony aresh own in Tables I and 11. As was mentioned previously, these recoveries were determined by using spiked samples as the standard of comparison. Hydride generation interference effects have been noted with the addition of a variety of different cations to solutions of the elements being determined. The average deviations of four determinations indicate the relative precision as determined with this technique. Such precisions are satisfactory, especially when the TABLE I STANDARD RECOVERIES OF ELEMENTS FROM WHOLE BLOOD Average deviation* Amount added/ Amount recovered, from mean, Element Pg % Mean, yo % Ge 0.4 91, 91, 84, 82 87 4 As 1 .o 100, 108, 94, 95 97 3 Se 2.0 99, 105, 100, 101 101 2 Sn 1 .o 113, 108, 119, 107 112 5 * Average deviation from the mean value of four replicate determinations. TABLE I1 STANDARD RECOVERIES OF ELEMENTS FROM ENRICHED FLOUR Average deviation* Amount added/ Amount recovered, from mean, % '% Mean, yo Element Pg As 1.0 105, 95,.91, 105 99 6 Se 2.0 77, 114, 97, 77 91 14 Sb 0.5 87, 84, 77, 84 83 3 * Average deviation from the mean value of four replicate determinations.January, 1979 TIN AND ANTIMONY I N COMPLEX SAMPLES 39 low detection levels and numerous manipulation steps are considered.When compared with the relative precision data obtained from aqueous solutions listed in Table 111, there is no noticeable increase in the scatter of results that could be attributed to the introduction of the complex samples. It is also important to note that the absolute masses listed do not reflect the final solution concentrations that are dealt with, or that could be dealt with. To illustrate this, a com- parison of detection limits in micrograms is listed in Table I11 along with the solution concentration, assuming that the analyte is generated from the customary 20 ml of solution. TABLE I11 DETECTION LIMITS FOR AQUEOUS STANDARDS Data taken from reference 6. Wavelength/ Detection Detection Relative standard Element nm limit/pg* limit, p.p.b.* deviation? Ge 303.9 0.003 0.15 3.2% at 1 p g As 193.7 0.007 0.35 6.5% at 0.25 pg Se 196.0 0.025 1.25 5.5% at 1 p g Sn 317.5 0.040 2.00 2.9% at 1 pg Sb 259.8 0.010 0.50 8.8% at 0.1 p g * Using a 20-ml sample volume and defined as 2 x the standard deviation of t Standard deviation of 10 determinations as emission peak areas divided by base-line noise. the mean emission peak area and converted to percentage.Values for arsenic and antimony in NBS orchard leaves are shown in Table IV, as deter- mined by the standard additions technique. It should be noted that the arsenic deter- mination is well within the expected range. The antimony determination gives lower results than the NBS values, but the deviations meet. The difference in the values could be attributed to the fact that the NBS value was obtained by neutron-activation techniques.Although a value for selenium was determined by the NBS, also using neutron activation, this element was not determined by these laboratories as the amount of leaves needed for each analysis became unreasonably large for digestion and subsequent determination with the standard additions technique. TABLE IV DETERMINATION OF ARSENIC AND ANTIMONY IN NBS ORCHARD LEAVES 1571 Arsenic content/pg g-l Certified ’ Sample Found value 1 9.3 2 9.4 3 8.3 4 8.9 Mean 9.0 & 0.4* 10.0 f 27 Antimony contentlpg g-l Found value A r 1 Certified 2.8 2.3 2.1 2.0 2.3 & 0.3* 2.9 f 3t * Average deviation from mean. t Two standard deviations “of entire range of observed results.” In summary, the results indicate that hydride generation coupled with microwave-induced plasma atomic-emission spectroscopy is a useful technique for analysis of complex samples. In fact, the low detection limits and the degree of reproducibility that the technique offers make it not only superior to techniques involving solution nebulisation, but the charac- teristic long dynamic linear ranges and potential for multi-element analysis, which the microwave-induced plasma offers, make it potentially more useful than the more widely used atomic-absorption - hydride generation techniques of analysis.40 ROBBINS, CARUSO AND FRICKE References 1. 2. 3. 4. 5. 6. 7. Rliyazaki, A., Kimura, A . , and Umezaki, Y., Analytica Clzim. Acta, 1977, 90, 119. E'iorino, J . A., Jones, J . W., and Capar, S. G., Analyt. Chem., 1976, 48, 120. Drinkwater, J . E., Analyst, 1976, 101, 672. Seimer, D. D., and Koteel, P., Analyt. Chenz., 1977, 49, 1096. Shaikh, A. U., and Tallman, D. E.. Analyt. Chern., 1972, 49, 1093. Fricke, F. L., Robbins, W. B., and Caruso, J. A., J . Ass. Off. Analyt. Chem., 1978, 61, 11s. Smith, A. E., Analyst, 1975, 100, 300. Received March 14tli, 1978 Accepted July loth, 1978